Currently how many of you are using Google Assistant to know Search for Your favorite movie, favorite song or any questions and might be using for many other things. Probably everyone, except pity users like me who are having bricks to hit hard people when you are angry with someone. Joke apart, Now Google is coming up with more powerful virtual assistant Google Duplex.
How Google Duplex is different from other assistant?
Current Virtual assistant voice seems more like robotic sound but as AI goal is to Virtualize everything that human can do just like human or better than Human. Duplex not only able to understand and process the natural language but to generate human like Sounds with being more smarter. As said by Yaniv Leviathan and Matias.
They don’t engage in a conversation flow and force the caller to adjust to the system instead of the system adjusting to the caller.
Hurdles in Path
As humans are more irrational and non monotonic, they can answer differently in different context. Some are fluent while talking while some stumbled, as an instance of conversation referred in Google AI Blog by their engineer:
“So umm Tuesday through Thursday we are open 11 to 2, and then reopen 4 to 9, and then Friday, Saturday, Sunday we… or Friday, Saturday we’re open 11 to 9 and then Sunday we’re open 1 to 9.”
How it Works?
It is build using TensorFlow and Neural Network subbranch Recurrent Neural Network
Recurrent nets are a type of artificial neural network designed to recognize patterns in sequences of data, such as text, genomes, handwriting, the spoken word, or numerical times series data emanating from sensors, stock markets and government agencies.
Deeplearning4j
For speech engine that sounds natural, they used use a combination of a concatenative text to speech (TTS) engine and a synthesis TTS engine (using Tacotron and WaveNet) to control intonation depending on the circumstance.
Resources:
Recurrent Neural Network – Deep Learning 4j
Google AI Duplex – Google Ai Blog
